Design of video streaming service quality control system using available bandwidth information Speaker : 吳靖緯 MA0G th International Conference on Advanced Communication Technology (ICACT), On page(s): , Feb. 2011

Outline Introduction Related works Proposed VQCS operating environment and functional architecture Performance analysis Conclusion 2

Introduction Preexisting Internet service and variance multimedia service such as IPTV(lnternet Protocol TV), VoIP(Voice over IP) and VoD(Video on Demand) are provided in converged network environment. QoS( Quality of Service ) control architectures recommended by standards institutes provide resource management and admission control only but do not consider service user's experience quality. 3

Introduction QoE(Quality of Experience) is the term which is used to describe how it is satisfied by subscribers to the provided service quality. QoE is very subjective in nature, it is very important that a strategy is devised to measure it as realistically as possible. This propose the system of interworking control between media quality and transport QoS parameters especially available bandwidth to enhance the service user's quality level. 4

Introduction Service and network providers can offer the enhanced service quality to users using the proposed VQCS(Video Quality Control System) system. It can adaptively control the data rate of stream according to network available bandwidth. 5

Related works A.QoS control architectures in NGN(Next Generation Network) Existing video quality control system measures network delay, jitter and packet loss rate using control packet like RTCP(RTP Control Protocol). Then, it analyzes the measured QoS parameters, and derives QoS control values to enhance the experienced video quality. 6

Related works Existing quality control system does not consider the available bandwidth, so it is irrelevant to control the quality of video streaming service. Because a video streaming service requires a large bandwidth, the lack of available bandwidth leads to packet loss. The real-time available bandwidth information management is required because packet loss causes serious video quality degradation such as jerkiness, block distortion and blurriness. 7

Related works B.The studies of link available bandwidth measurement 1.Available bandwidth measurement method using proving packets The existing available bandwidth measurement method is classified into the single-hop gap model based packet train measurement techniques and the self induced congestion techniques. The representative single-hop gap model has IGI(Initial Gap Model)and PTR(Packet Transmission Rate). 8

Related works It is the technique which observes the moment when the speed of being observed in the receiving side becomes smaller than the speed of the packet which a person transmits while having TOPP, a pathChirp, Pathload, and etc. The congestion induction technique enhancing the transfer rate of a packet with a queuing and measures the available bandwidth. 9

Related works 2.Available bandwidth measurement method using total length field in IP header The total length field within IP packet header is used in order to measure the per link available bandwidth. Each node extracts the information of the total length from the IP packet imported to router for measurement time and calculates the amount of the packets occupying a link through the cumulative summation of total length values. 10

Related works C.Study on the SVC(Scalable Video Coding) considering network environment and terminal performance SVC is a standardized video coding technology by JVT(Joint Video Team) which is composed of the ISOIIEC MPEG(Moving Picture Expert Group) and ITUT VCEG(Video Coding Expert Group). SVC considers image resolution and data rate to control video quality using spatial and temporal scalability. 11

Proposed VQCS operating environment and functional architecture A.VQCS operation environment Figure 1 shows the operating environment for VQCS that is proposed to control video quality using available bandwidth. 12

Proposed VQCS operating environment and functional architecture B.Functional architecture of VQCS Figure 2 shows the proposed VQCS system architecture which can actively enhance user's experience video quality by controlling stream data rate according to available bandwidth. VQCS is composed of 5 kinds modules, 2 exterior interfaces(Qs and QE) and 4 interior interfaces (Qa, Qf, Qm and Qr). 13

Proposed VQCS operating environment and functional architecture 14

Proposed VQCS operating environment and functional architecture C.VQCS operating procedures 15

Performance analysis construct test network as figure 4 to experiment the proposed VQCS performance. 16

Performance analysis This paper use the following commands to control network delay, jitter, packet loss and bandwidth in router based on linux(kernel 2.6 distribution support). #tc qdisc add dev eth0 root netem delay 100ms #tc qdisc change dev eth0 root netem delay l00ms 10ms #tc qdisc change dev eth0 root netem loss 0.1 % #tc qdisc add dev eth0 root handle 10: cbq bandwidth 5Mbit avpkt

Performance analysis This paper compare 300 original images and 300 test images passed through various network environments like figure 1 using full-reference method. 18

Performance analysis If video stream with lower 900Kbps data rate is transmitted in same network environment, the quality of video stream service can be improved by preventing packet loss although image resolution is decreased to of SSIM score as shown figure 5-(b). 19

Performance analysis Video streaming with 2.5Mbps data rate has high packet loss rate in 2Mbps available bandwidth network, so serious degradation occurs as shown figure 5-(c). 20

Performance analysis Figure 6 shows the improved video quality by using the proposed VQCS scheme because VQCS can adaptively control streaming data rate according to network resource status in order to prevent packet loss. 21

Conclusion Various multimedia services such as video, voice and gaming service can be provided by IPTV, and these services require large amounts of bandwidth. At this time, video quality degradation like video jerkiness, block distortion and blurring is caused when network available bandwidth is insufficient. So, we periodically calculate the amount of the packets in link and measure available bandwidth by using total length field in IP header at terminal. 22